Justus, Josephus, Agrippa II and his Coins

Justus, Josephus, Agrippa II and his Coins* Nikos Kokkinos An appendix to my book on the Herods dealt with the complex subject of the date of Agrippa II’s death.*1 A double critique of my view by Christopher Jones and Alla Kushnir-Stein has appeared in this journal.2 Jones claims to be making a contribution by means o f ‘Greek philology and Roman imperial prosopography’, tools presumably ‘in­ sufficiently exploited’, and declares that while the revised edition o f Schürer is out of date, my discussion is ‘questionable on many points’. So it is necessary to restate my case here. Had the numismatic model proposed by Kushnir-Stein been right, it would simply have made one aspect of my previous discussion, like that of Schürer (and of all numismatists involved), out of date — nothing more. But the model is not right as it stands, and in any case it does not affect my chronology. The appendix first exposed the weak assumptions behind the conventional date of pre-CE 93 for the death of Agrippa II, which are based on an interpretation o f Josephus. These assumptions will be mentioned again later in response to Jones. The appendix then listed evidence — literary and documentary — which contradicted the conventional date: the testimony of Justus o f Tiberias; the inscription of Archieus as interpreted by me; the lead weight from Tiberias as read by Qedar and interpreted by me; and the coins o f Agrippa II as interpreted by Dan Barag (followed by Meshorer and Burnett e t a l.). The documentary evidence will be re-assessed last in response to Kushnir-Stein. Justus The crucial testimony of Justus, placing the death of Agrippa II in CE 100, was included in his lost work [Χρονικὸ ν] Ίουδαἰ ων Βασιλἐ ων τῶ ν ἐ ν το ῖ ς Στἐ μ μ ασιν ([ C h r o n ic le ] * 1 2 Scholars who must particularly be thanked for discussion are Dan Barag, Theodore Buttrey, Louis Feldman, Tomas Hägg, Brian Jones, Fergus Millar, Daniel Schwartz, Joseph Sievers, Warren Treadgold and Nigel Wilson. The following abbreviations are used in this article: B M C R E = Η. Mattingly, C oins o f the R o m a n E m p ir e in the B ritish M useum , vol. 1 (London, 1923); C D P A = Α. Spijkerman, The C o in s o f the D eca p o lis a n d P ro vin cia A ra b ia (Jerusalem, 1978); K H D = Ν. Kokkinos, The H ero d ia n D yn a s ty (Sheffield, 1998); R P C = Α. Burnett, Μ. Amandry & I. Carradice, R o m a n P ro vin c ia l C oinag e, vol. 2 (London/Paris, 1999); S V M = Ε. Schürer, The H is to ry o f th e Jew ish P eo p le in th e A g e o f Jesu s C hrist, vol. 1, rev. by G. Vermes & F. Millar (Edinburgh 1973); T JC = Y. Meshorer, A T reasury o f J e w ish C oins (Jerusalem/New York, 2001). K H D 396-400. C.P. Jones, ‘Towards a Chronology of Josephus’, S C I 21 (2002) 113-21 (henceforth ‘Jones’); Α. Kushnir-Stein, ‘The Coinage of Agrippa ΙΓ, 5C /21 (2002) 123-31 (henceforth ‘Kushnir-Stein’). S cr ip ta C la ssica Isr a elica vol. XXII 2003 pp. 163-180 164 JUSTUS, JOSEPHUS, AGRIPPA II AND HIS COINS o f the K in g s o f the Je w s w ho a re in th e G en ealo gica l T a b les)? A brief review written around CE 845 has been preserved in the Βιβλιοθὴ κη o f Photius (codex 33). The main summary reads: Ἀ ρ χ ε τ α ι δὲ τῆ τοὺ ὲ βδὸ μ ου μ βασιλεῦ σιν, ὅ ς ὲ τι μ ᾶ λλον ὺ πὅ κατἐ ληξεν. ς ἱ στοριας ἀ πὸ Μωῦ σἐ ως , καταλῆ γει δὲ ἕ ως τελευτῆ ς ὲ ν τῶ ν ἀ πὸ τῆ ς οἰ κἰ ας Ἡ ρῷ δου, ὺ στἀ του δὲ ἐ ν το ῖ ς παρἐ λαβε μ ὲ ν τῆ ν ἀ ρχῆ ν ἐ πὶ Κλαυδἰ ου, ηὺ ξῆ θη δὲ ἐ π'ι Νἐ Οὺ εσπασιανοὺ , τελευτᾷ δὲ ἔ τει τρἰ τῳ Τραιανοῦ , οὐ καῖ Ἀ γ ρ ιπ π α Ίουδαἰ ων ρωνος καῖ ῆ ιστορἰ α He begins the history from Moses and concludes with the death o f Agrippa, the seventh of those from the House o f Herod and last to reign among the Jews, who received the authority during the time of Claudius, increased it during the time of Nero and even more under Vespasian, and died in the third year of Trajan, when the history also ends.4 While this evidence has been readily accepted by many scholars o f older generations,5 in recent decades there has been an array o f attempts to reject it. Few other such testimo­ nies can have been ignored, sidestepped, amended, mangled, mistranslated, and misin­ terpreted as much as this. Jones (116), avoiding individual evaluations, offers what he thinks to be a solution: ‘Various ways around it have been proposed, some more plausi­ ble than others. If they are rejected, the only way to reconcile Photius’ statement with the supposition that Agrippa had already lost his kingdom is to emphasize the word “died”, and to assume that he lived on as a titular king’. But this will not do. The ‘supposition’ is only that, and the emphasis is o f course on the word ‘authority’ which was clearly essential in a chronicle of royal succession — Agrippa’s authority must have ended with his death in the third year of Trajan. The most usual ‘ways around’ this evidence since the 1960s, apart from doubting whether Justus’ work was available to Photius (despite the B ib lio th e c a 's repeated The meaning is not clear, but it seems to suggest the existence of royal stem m a ta , which may only refer to the genealogical king-lists in the Bible. Perhaps the brief co m m en ta rio li d e sc rip tu ris mentioned by Jerome ( Vir. Illu st. 14) as written by Justus is the same work on biblical kings. On Justus (P IR 1 1 872; F G r H 734) see, C.R. Holladay, F ra g m en ts fr o m H el­ len istic J ew ish A u th o rs, vol. 1 (Chico, CA, 1983) 371-89 with bibliography. The text is from the edition o f R. Henry (Paris 1959-77), while the translation is mine after checking the French of Henry, the Latin of Α. Schott (P G 103), and the modem Greek of S. Euthymiades (Βιβλιοθῆ κη ὅ σα τῆ ς Ίστορἰ ας : Ἀ νθολογἰ α [Athens, 2000]). Since codex 33 is not included in Ν. Wilson (P ho tiu s, the B ib lioth eca : A S e lectio n [London, 1994]), I pre­ sume that the only other English translation is in S.J.D. Cohen (,Jo sep hu s in G a lilee a n d R o m e [Leiden, 1979] 142); cf. part translation in D.A. Barish (‘The A u to b io g ra p h y of Jo­ sephus and the Hypothesis o f a Second Edition o f his A n tiq u itie s ’, H T R 71 [1978] 71). For Photius see primarily Τ. Hägg, P h o tio s als V erm ittler a n tiker L ite ra tu r (Uppsala, 1975); W.T. Treadgold, The N a tu r e o f the Bibliotheca o f P h o tiu s (Washington, DC, 1980); J. Schamp, P h o tio s h isto rien d e lettres (Paris, 1987); for an exhaustive bibliography see G. Dragas, ‘Towards a complete Bibliographia Photiana’, Έκκλησἰ α καὶ Θεολογἰ α 10 (1989/90) 531-669. For an excellent historiographical essay on this problem, see D. Schwartz, ‘Texts, Coins, Fashions and Dates: Josephus’ Vita and Agrippa II’s Death’, in S tu d ies in the J e w is h B a c k­ g r o u n d o f C h ristian ity (Tübingen, 1992). NIKOS KOKKINOS 165 formula Άνεγνωσθη!), have been to question the reading of the given date, or present it as confused, and then launch an attack a d h om inem . How could a ninth-century Christian writer have transmitted information o f value? Any attempt to give priority to Photius can only be unjustified, as its ‘only purpose’ would be ‘to defend the honor of a tenth (sic) century patriarch’.*6 He surely must have erred in copying Justus. ‘If the wrong date were inserted by Photius, such an error would be unsurprising’.7 Or, in any case he should simply have ‘confused the date of Justus’ publication (s ic )... with the date of Agrippa II’s death’.8 Clearly the best work of what is regarded as the best scholar in Byzantium had not been read carefully, and modem specialists on Photius unfortunately were not consulted. To disbelieve the specific information in codex 33 is extremely difficult given its context. Photius must have read the particular work o f Justus — at least its beginning and end, to which there are exact references. His comments on the author’s style also support this view. Photius is not accused of claiming to have read books that he had in fact not read; and there is one codex (268) in which he says explicitly that he had not been able to read the author in question. The possibility of misquotation is very low, as he is known to be correct in cases o f this kind, and there could be no motive for him to invent such a detail. Distortion due to scribes is also very unlikely. The text is based on two MSS only, which look as if they are descendants from an original master copy. The key sentence which is missing in MS Μ is added in the margin in MS A, leaving no doubt that it was written in the archetype. The inevitable conclusion is that the burden of proof certainly lies with anyone who would argue that Photius is wrong. Such a conclusion makes it almost unnecessary to probe further into codex 33, unless it is to explain a final misunderstanding. It was once suggested that since in Jerome’s D e Viris Illu strib u s the entry on Justus (14) is followed by that on Clement of Rome (15), which contains the words o biit tertio T ra iani a n n o , Photius’s τελευτᾷ δὲ ἔ τει τριτῳ Τραῖ ανοὐ may have been transferred by mistake from this source.9 This interesting but superficial suggestion was soon to be taken enthusiastically up, and even repeated recently as a solution.10 But it cannot stand up to critical scrutiny. The 280 numbered codices o f the B ib lio theca (which in fact describe some 386 different books) show that Photius either had no access to Latin literature or, in any case, made no use of it. Photius’ acquaintance with Jerome was merely at third hand. Jerome’s work (CE 6 7 8 9 10 Cohen (n. 4) 180. Τ. Rajak, ‘Josephus and Justus of Tiberias’, in L.H. Feldman & G. Hata (eds), Jo sep hu s, Ju d a ism & C h ristian ity (Detroit, Mich., 1987) 93, n. 10 = The J ew ish D ia lo g u e w ith G reece & R o m e (Leiden, 2001) 180, n. 10. B.Z. Wacholder, E u p olem us: A S tu d y o f Ju d a e o -G re ek L ite ra tu re (Cincinnati, 1974) 302; followed by Barish ([n. 4] 72) and by S. Mason (F la viu s J o se p h u s , Vol. 9 : L ife [Leiden, 2001] xviii), but the idea goes back to S. Basnage in 1707 (see Schwartz [n. 5] 250, n. 21). Not only is the Greek being abused here but these scholars cannot have troubled to compare codex 33 with other codices in the B ib lio th eca (cf. comment in Schwartz [n. 5] 248, n. 16). The Greek translation of Jerome by ‘Sophronius’ (ed. Fabricius), actually reads τελευτᾷ τῷ τρἰ τῳ ἐ νιαυτῷ Τρά ΐανοϋ. For the suggestion see, S V M 1: 482, n. 47 (1); cf. Τ. Rajak, ‘Justus of Tiberias’, C Q 23 (1973) 362, n. 2 = The J ew ish D ia lo g u e w ith G reece & R o m e (Leiden, 2001) 166, n. 14; Barish (n. 4) 72; Mason (n. 8) xviii-xix. 166 JUSTUS, JOSEPHUS, AGRIPPA II AND HIS COINS 392) had been translated into Greek by a ‘Sophronius’ — probably his friend mentioned in Vir. Illust. 134. ‘Sophronius’ had been a source of Christian biographies for a ‘bio­ graphical dictionary’ o f the ninth century, used upon occasion by Photius and later by the S uda . This ‘dictionary’ (usually called ‘Hesychius Epitome’) was based on the sixthcentury O n o m ato log us of Hesychius, which contained only pagan biographies. The important thing to understand is that while ‘Sophronius’ was arranged roughly chrono­ logically and the O n o m ato log us precisely thematically, the ‘Hesychius Epitome’ was arranged completely alphabetically." So the Latin Christian biographies of Jerome (which had the entry on Justus followed by that on Clement o f Rome), being translated into Greek by ‘Sophronius’, lost their chronological order to the a lp h a b etic a l ‘Hesychius Epitome’ known to Photius. The entry on Ίοΰστος ( Vir. Illu s. 14) would have been followed by that of Ίππό λυτος of Rome ( Vir. Illu s. 61), and then by Ίωά ννης the Evangelist ( Vir. Illu s. 9) and so on. The entry on Clement/Κλὴ μ ης of Rome ( Vir. Illu s. 15) could not have followed that of Justus, and thus there was no way for Photius to be mistaken.1112 But what should make this suggestion finally obsolete is the fact that Photius, in the case o f Justus, does not borrow any biographical details from the ‘Hesychius Epitome’. While some information may have been obtained from the work o f Justus itself, the main biographical details come from Josephus — as Photius admits by citing him. A further source which displays points of contact with the particular biography is Philostorgius (S uda , s.v. Φλέ γων), clearly known to Photius (codex 40). But in thinking of the order o f entries, and in sharp contrast to the suggestion about Jerome, yet another source can be revealed in the B ib lio th eca . First it is necessary to quote Treadgold on the method used by Photius in writing: As he was dictating the first part of the B ib lio th e c a ... Photius kept a manuscript o f the ‘Hesychius Epitome’ by his side and looked up its alphabetical articles on approximately twenty of the authors he was reviewing. More often, he looked for autobiographical mate­ rial in the authors’ own books. He might also use a book he reviewed in one codex as a source for the literary history of another codex. Once or twice his use of additional sources may have affected the order of the codices. Thus, Photius treats Theopompus after Pamphila, who may have excerpted biographical material on Theopompus; he describes works by Clement of Rome after works by Clement o f Alexandria, who would have preceded Clement o f Rome in the alphabetical ‘Hesychius Epitome’, consulted by Photius on both writers.13 11 12 13 See G. Wentzel, ‘Die griechische Übersetzung der Viri in lu stres des Hieronymus’, T U 13.3 (1895) 57-63. Note that in the Latin context o f Jerome the entry on Clement (15) which includes ‘the third year o f Trajan’, is followed by the entry on Ignatius (16) which refers to ‘the eleven year of Trajan’, while the entry before Justus (14), that o f Josephus (13), mentions ‘the fourteenth year o f Domitian’. The latter date also contradicts our received text of Josephus (Ant. 20.267)! Treadgold (n. 4) 66. Note that Clement of Alexandria and Clement of Rome are in codices 109-13. NIKOS KOKKINOS 167 Since we know that Justus’ work was almost certainly used as a source by Julius Africa­ nus,14 it should come as no surprise that Africanus is registered immediately after Justus in the B ib lio th eca (codex 34). If we note with Treadgold that Photius’ use of additional sources may occasionally ‘have affected the order of the codices’, then it becomes evident that Photius could have obtained biographical information for Justus also from Julius Africanus.15 The CE 100 date for the death of Agrippa II copied by Photius from the work of Justus, the disgraced but very knowledgeable (Jos. L ife 40) secretary and court historian of Agrippa II, must be correct. Daniel Schwartz rightly stressed that ‘... the only explicit statement anywhere regarding the year of Agrippa II’s death is offered by Photius, in the course o f his account of Justus of Tiberias ... we should demand strong reasons to set the date aside’.16 The arguments which led Emil Schürer over a century ago to the important conclusion, that ‘there is no reason for doubting the correctness of this statement’,17 should have never been discarded. Josephus The date of W ar, the earliest work of Josephus, is not directly relevant to the present discussion and thus can be left out of account here.18 Concerning the dating of A n tiq u i­ ties, the question of the identity of Epaphroditus, Josephus’ principal patron,19 may also be sidestepped due to lack o f decisive evidence.20 One must then ask, does Josephus in 14 15 16 17 18 19 20 Η. Geizer, S extu s Ju liu s A frica n u s u n d d ie b yza ntin isch e C h ro n o g ra p h ie (Leipzig, 1880) 246-65. Cf. S. Bowman, ‘Josephus in Byzantium’, in L.H. Feldman & G. Hata (eds.), Jo seph us, Ju d a ism , a n d C h ristian ity (Detroit, Mich., 1987) 379, n. 24. Schwartz (n. 5) 272-3. Ε. Schürer, A H isto ry o f the Jew ish P eo p le in the Tim e o f J e su s C hrist, vol. 1 (Edinburgh, 1890) 205-6. la m more or less in agreement with Jones’ dating of the W ar, but it must be noted that his rejection of Seth Schwartz’s proposal of a very late date for Book 7, is partly based only on one observation o f mine, but more have been given (cf. K H D 199, n. 98 & 252, n. 25). Α date as early as 75, appropriately coinciding with the completion o f the Temple of Peace (Dio 66Ἰ 5.3), and when Agrippa II was in Rome (which would have enabled him to acquire a copy — A p . 1.51), is possible for an ea rliest version. In such a case W ar Τ Λ 58-62 (de­ scribing the temple) would be a later addition, but not later than 79-81. Jones, p. 113, calls: ‘Josephus’ principal patron, Agrippa II (s ic )'. Agrippa II may have had a dialogue with Josephus regarding the narrative in W ar, over ‘62 letters’ evidently not al­ ways agreeable (L ife 364-6; cf. A nt. 16.187), but he was certainly not Josephus’ principal patron — not then, not later. Agrippa II was Justus’ principal patron until the latter’s banishment (L ife 355-6). As Jones (114-5) admits the name is common in the Flavian period and not worth arguing for, even though the best candidate is the grammarian Μ. Mettius Epaphroditus (P IR 2 Μ 563; cf. L.H. Feldman, F la viu s Jo sep h u s, Vol. 3: A n tiq u ities 1-4 [Leiden, 2000] 5, n. 9), who was still alive under Nerva. Note significantly that he is the only Epaphroditus men­ tioned in the S u d a (s.v.), and since his time is said to have overlapped with that o f the grammarian Ptolemy the son o f Hephaistion, who flourished under Trajan and Hadrian 168 JUSTUS, JOSEPHUS, AGRIPPA II AND HIS COINS A ntiq u ities provide a different, a b so lu te date from that of Justus regarding the death of Agrippa II? The answer is emphatically no. Current thought interprets certain passages (17.28; 20Ἰ 41-7, 211-4) as meaning that Agrippa II had died before the completion of A n tiqu ities, which is expressly said (20.267) to have occurred in CE 93/4. But none of these passages permits an exclusive interpretation. A nt. 17.28 can at best show that Batanaea was lost to Agrippa II by 93, and this is confirmed by inscriptions indicating that Auranitis was annexed by 96 and Trachonitis by 96/7.21 There was nothing unusual in client kings gradually gaining or losing territories, or even being transferred to differ­ ent kingdoms altogether. For example, Armenia Minor, ruled by Agrippa IPs cousin Aristobulus III since 54, was taken away from him around mid-71, when he was trans­ ferred to the kingdom of Chalcis (where we find him in 72/3). Sohaemus of Emesa would also have lost Sophene, which was presented to him at the same time as Aristobulus’s appointment. And while Armenia Minor had previously been ruled (38 to before 54) by Cotys son of Cotys of Thrace, part of Cilicia was granted to Polemo II of Pontus in 41, who lost Pontus in 63, while other parts had been left periodically in the hands of kings from Cappadocia and Commagene.22 The ending o f A n t. 17.28 (καὶ τά δε μ ὲ ν ἤ καιρὸ ς ά κριβῶ σομ αι προῖ ό ντος τοΰ λό γου) also suggests the lateness of the informa­ tion about Batanaea, as Josephus gives a promise to relate events after the end o f the rule o f Agrippa II, which ‘apparently’ he did not keep. This promise, together with others, Josephus decided later (20.267) to fulfil in a future, updated version o f his history — a plan which he seems to have abandoned. But as argued by Altshuler in the case of ‘unfulfilled’ promises for a future work on the Mosaic Law, Josephus would have abandoned this plan only in favour of going back and making at least some brief additions to the A n tiq u itie s,23 As to A n t. 20.141-7 & 211-4, the hostile comments about Agrippa II and his family by no means require him to be dead. Josephus is often inconsistent in his remarks: for example Agrippa II is praised both earlier and later in the same book (20.9, 12, 135, 328). Antipater Ι, Agrippa II’s great-great-grandfather, is labelled a ‘troublemaker’ (στασιαστῇ ς ) in A n t. 14.8, but in A nt. 14.101 ‘a man of good sense’ (συνετό ς ). Numer­ ous examples can be adduced. Such inconsistencies are explained by uncritical copying (,Sud a s.v.), Epaphroditus would have died early in the reign o f Trajan (cf. Schwartz [n. 5] 21 22 23 267). See K H D 338. Α co h ors p r im a F la v ia C a na th eno ru m (C IL 8.2394-5 e t a l.) may either have been recruited from Canatha in Auranitis during the Jewish War and incorporated later into the Roman army, or else recruited by 96 after the territory’s annexation (cf. Μ.Μ. Roxan, R o m a n M ilita ry D ip lo m a s 1985-1993 [London, 1994] 273, n. 4). In fact coins o f Canatha show that the city was not under Agrippa H’s jurisdiction by 94/5 (see below n. 52). See K H D 312 (Aristobulus III), 249 (Sohaemus), 311 (Cotys), 381-2 (Polemo II), 252-3 (kings o f Cappadocia and Commagene). D. Altshuler, ‘The Treatise ΠΕΡΙ ΕΘΩ Ν ΚΑΙ Α1ΤΙΩ Ν O n C ustom s a n d C auses by Flavius Josephus’, J Q R 69 (1978/9) 226-32; cf. Η. Petersen, ‘Real and Alleged Literary Projects of Josephus’, A J P 79 (1958) 259-74. It is then even possible that the brief statement about the Romans in Batanaea was added later to A n t. 17.28, as a part fulfilment o f the ‘remaining’ promise given by Josephus. Cohen ([n. 4] 179, n. 248) allows as plausible that the entire section 17.23-31 is a later addition. NIKOS KOKKINOS 169 from different sources,24 and/or by later revisions, following the well-known argument of two editions for the A n tiq u itie s 25 This old argument does not need an introduction here, and although current conventional thought would wish it to be consigned to oblivion, it is simply too strong for that. An interesting paradox is found in the revised Schürer. The revisers who have on the one hand rejected Schürer’s own opinion concerning the validity of the date provided by Justus/Photius, on the other favoured the argument of two editions for the A n tiq u itie s.26 The time has certainly come for a fresh investigation of the matter as also suggested recently by Sievers: ‘The whole question of a second edition seems to be worth taking up again, re-studying also, but not exclusively, the two passages at the end of the A n tiq u itie s and the L ife . Here one would need to keep in mind the peculiarities o f producing a ‘second edition’ o f an ancient work in manuscript form’.27 A n t. 16.187 is an important passage in which Josephus clearly alludes to the fact that Agrippa II was still a live a little before 93/4. The commentary provided by Jones (117), based on the frequently erroneous Marcus/Wikgren volume of the Loeb edition, is mis­ leading. The sentence πολλοὺ ς μ ὲ ν τῶ ν ἐ γγονῶ ν τῶ ν ἐ κεἰ νου βασιλεὐ οντας ἔ τι δ ι ’ ἐ ντροπῇ ς ἔ χοντες , can only be understood as ‘and although we have respect for many o f his grandchildren (i.e. descendants) who are still reigning’ — with ‘his’ referring to Herod the Great in the general context.28 It is not possible to connect ‘his’ with Ἀ σαμ ω ναῖ ος mentioned earlier in the paragraph, for that does not make sense. No Hasmonaeans were reigning at the time Josephus was writing. But even if the intention was to call ‘Hasmonaeans’ the particular descendants of Herod, stressing their remote connection to his Hasmonaean wife Mariamme I, the result is exactly the same. The ‘many’ kings reigning shortly before the completion of the A n tiq u itie s in any case be­ longed o n ly to the lineages o f Mariamme’s sons (Alexander I and Aristobulus I). This is 24 25 26 21 28 See D. Schwartz, ‘ΚΑΤΑ TOYTON ΤΟΝ ΚΑΙΡΟΝ: Josephus’ Source on Agrippa II’, JQ R 72(1982) 241-68. Primarily R. Laqueur, D e r jü d is c h e H isto rike r F la viu s Jo se p h u s (Giessen, 1920) and Η. St. J. Thackeray, Jo sep h u s the M a n a n d the H isto ria n (New York, 1929); but selectively see Α. Momigliano, ‘Josephus as a Source for the History o f Judaea’, in C A H 10 (1934) 886; G .C. Richards, ‘The O p p o sitio n of Josephus’ “Antiquities’” , C Q 33 (1939) 36-40; Altshuler (n. 23); Schwartz (n. 24); J. Sievers, ‘Josephus, First Maccabees, Sparta, the Three “Haireseis” — and Cicero’, J S J 32 (2001) 241-51; cf. H.W. Attridge, The In terp re ta tio n o f B ib lica l H is to r y in the Antiquitates Judaicae o f F la viu s Jo sep h u s (Missoula, Mont., 1976) 52, n. 2; L.H. Feldman, Jo sep h u s a n d M od ern S c h ola rship , 193 7-1 98 0 (Berlin, 1984) 837-8. Compare S V M 1:54 with 481-2. While one may be certain that pp. 481-2 were written by Fergus Millar, the preface to the book (vi) reveals that p. 54 was revised by Tessa Rajak (cf. G. Vermes, ‘How the New Schürer Came into Being’, in Α. Oppenheimer [ed.], Jü d isc h e G esch ich te in h ellen istisch -rö m isch er Z e it [Munich, 1999] 6). However, Rajak was never in favour o f the two editions argument ([n. 10] 361 = 165; Jo sep h u s [London, 1983] 237-8). Sievers (n. 25) 248. I am grateful to Louis Feldman for discussing the possibility of a ‘sec­ ond edition’ to both A n tiq u ities and Life. Cf. also the old translation of W. Whiston, The W orks o f F la viu s J o se p h u s, vol. 2 (Edin­ burgh, 1826) 403, and the modern Greek translation of the Kaktos edition, Ίωσηπος : Ἀ π α ν τα , vol. 13 (Athens, 1997) 245. Josephus refers many times to ‘his’, ‘hers’ and ‘theirs’ without specifying that he means the Herodian family and their Idumaean race (e.g. A n t. 14.300, 379) 170 JUSTUS, JOSEPHUS, AGRIPPA II AND HIS COINS surely good evidence for the reigning status shortly before 93/4 of Agrippa II, Aristobulus III o f Chalcis and Alexander IV o f Cilicia.29 Also the fact that Josephus in the fol­ lowing sentence reveals that he has ‘provoked those very persons to anger’, links well with one of the letters sent to Josephus by Agrippa II (L ife 366), in which the king politely invites the historian to be informed about much evidence ignored by him. The ‘anger’ o f the king further indicates that Josephus’ previously published work contained some unfriendly passages about the Herods, despite the fact that Agrippa II was still a live (e.g. W ar 1.196, 313, 317, 477-8, 521; 2.84-5 etc. — cf. discussion above on Book 20 o f A n t.). A nt. 18.128 is to be understood in a similar way — only here different rhetoric re­ quired the descendants of Herod to be presented as ‘a few’ (which they actually were) rather than ‘many’ (16.187 above), εἴ γε ἐ ντὸ ς ἐ κατὸ ν ἐ τῶ ν δ ι’ δλου30 συνέ βη πλῇ ν ό λἰ γων, πολλοὶ δ ’ ἦ σαν, διαφθαρἤ ναι τοὺ ς Ήρῶ δου ά πογό νους . ‘For it happened that within a hundred years overall, the descendants of Herod, who were many, perished with the exception o f a few’. Counting from the death o f Herod (4 [or 5/4] BCE) inclu­ sively, we have 98 years to 93/4 when A n tiq u itie s was completed (or less at the time of writing Book 16). This indeed is ‘within’ (i.e. ‘inside’) 100 years, when only ‘a few’ of Herod’s descendants were still active — Agrippa II, Aristobulus III and Alexander IV. The only allusion to Agrippa II’s death in Josephus is L ife 359,31 and the conven­ tional thought here assumes that since the L ife seems to have been written as a supple­ ment to the A n tiq u itie s (20.266), the whole project was completed and published in 93/4 (20.267). But there are serious problems with this assumption. L ife 336-67 is a παρἐ κβασις (367), known as the ‘great digression’,32 and strongly suspected of being a later addition. Whether the L ife was an original supplement to the A n tiq u ities depends on whether the conclusion 20.259-66 was itself original. For why then would there be a second conclusion immediately after 20.267-8? Further, what was really completed in 93/4 was the A n tiq u ities, not the L ife, and therefore neither need have been published at this date. In fact, given the years of Terror o f Domitian’s reign (93-96), it is very unlikely that the work circulated before the accession o f Nerva. Moreover, since A p io n 1Ἰ reveals that the A n tiq u ities had already appeared (presumably together with the L ife), it is extraordinary to find Josephus in A p io n 1.51 referring to Agrippa II as ‘the most wonderful (θαυμ ασιῶ τατος ) King Agrippa’. How could this statement, which is supposed to be later than Book 20 of A n tiq u ities, which is supposed to be hostile to 29 30 31 32 See K H D 213-5 (Mariamme I’s sons) 338 (Agrippa II, his cousin Aristobulus III, and his cousin once removed Alexander IV). Jones’ comments about Alexander IV (117 & n. 18) can safely be ignored, including his reference to Julia Crispina — see K H D 250-8 and 2934. The preferred reading o f ἐ ξὸ δου (in the Loeb edition) is not better than that chosen here, which also avoids the emendation περιὸ δου of Richards and Shutt. It is only an allusion because while for Vespasian and Titus the word used is ζῶ ντων (‘when they were alive’) the word for Agrippa is περιὸ ντος ἔ τι (‘when he was still around’) — which strictly speaking cannot exclude the possibility of him ‘riot being around as a k in g ’, but still alive. Nevertheless, the allusion to him being dead is strengthened immediately after (.Life 360). R.B. Motzo, S a g g i d i sto ria e lettera tu ra G iudeo-EU enistica (Florence, 1924) 222; cf. Cohen (n. 4) 114-20. NIKOS KOKKINOS 171 Agrippa, who was supposed to be dead, have been written? So the conventional view is far too simplistic and must be dropped. In L ife 360, immediately after the allusion to the death of the king (and still within the ‘great digression’), Josephus criticizes Justus’ delay in publishing his work on the Jewish War. He says that Justus wrote it when Vespasian, Titus and Agrippa II (mentioned before in the text) were still alive, but waited for ‘twenty years’ (εἵ κοσιν ἐ τῶ ν) until n o w when these people cannot testify to its accuracy, ‘not being with us any longer’ (οὐ κέ τ’ εἰ σ ιν μ εθ’ ῇ μ ῶ ν). Conceivably in a manner o f speech, the figure given is a round number. Since Vespasian died in 79 and Titus in 81, the publication o f Justus’ work is placed by Josephus roughly at the end o f the century, and thus a date immediately after the death of Agrippa II early in 100 is in good order. This agrees with the placing of Justus’ published life (συγγραφεὺ ς ἐ γνωριζετο) at the beginning o f the reign of Trajan by Georgius the Syncellus (ed. Dindorf I, 655). Georgius’ source was Julius Africanus who based his work on Justus himself — which takes us back to the same source. Jones (118-20) sees some difficulty with L ife 428-9 (outside the ‘great digression’), and believes (n. 22) that I have misunderstood the Greek in connection with Domitia of whom Josephus says that διετἐ λεσεν εὐ εργετοὐ σἀ μ ε. I have suggested that the phrase implies that Domitia continued to help Josephus after the death of Domitian. But apart from Liddell & Scott, Jones should have considered Rengstorf under διατελέ ω, where he would have found the meaning of ‘to continue, go on (from this time on, for all time)’.33 We know that Domitia survived Domitian, and continued to use his name unashamedly ( C IL 15.548a-9d).34 In conclusion so far, Justus offers an absolute date for the death o f Agrippa II while Josephus does not. Interpretations involving different passages and the whole structure o f the A n tiq u ities and L ife merely assume that Josephus indicates an earlier date than that given explicitly by Justus. But this is not proved or at all desirable. Fresh interpretations suggest that Josephus can be in perfect agreement with Justus. On the main issue o f the relationship between A n tiq u ities and L ife, I remain confident in my suggestion in the original appendix: although the A n tiq u ities was completed in 93/4, it would not have been published until later, or if it was, it would not have incorporated the L ife in the first edition, or if it did, the L ife was re-edited at a later stage. 33 34 K.H. Rengstorf, A C om p lete C o n co rd a n ce to F la viu s Jo sep h u s, vol. 1.2 (Leiden, 1973) 475; cf. ΡὈ . Dorbarakis, Ἐ πιτομ ον Αεξικὸ ν τῆ ς Ἀ ρ χ α ἰ α ς Ἐ λληνικῆ ς Γλῶ σσης (Athens, 1975) 225. Even Mason’s commentary on Josephus’ L ife ([n. 8] 172) had to admit partly on this point: ‘Kokkinos makes the novel argument... there is no grammatical problem.. Υ I am grateful to Fergus Millar, Louis Feldman and Brian Jones for discussion on this point. Α few Josephan examples will suffice: A n t. 10.215: διὸ μ ετά ταῦ τα πά σης ά ξιοὺ μ ενοι π α ρ ’ αὺ τοῦ τιμ ῆ ς διετἐ λουν (‘and so they continued thereafter to be held worthy by him o f the highest honour’); A n t. 15. 180: καὶ πά σας τά ς τιμ ά ς ά πολαβῶ ν ἔ τη τεσσαρά κοντα διετἐ λεσεν ἐ ν αὺ ταῖ ς (‘he received all his honours back and continued to enjoy them for forty years more’; L ife 423: καὶ τιμ ῶ ν διετἐ λει μ ἐ χρι τῆ ς ἐ κ τοῦ βΐου μ εταστά σεως ('he continued to honour me up to the time o f his departure from this life’). On Domitia (P IR 2 D 181 ; P F O S 327) see, B.W. Jones, The E m p e ro r D o m itia n (London, 1993) 33-8; cf. Ε Ἀ . Hemelnjk, M a tro n a D o cta (London, 1999) 299, n. 81. 172 JUSTUS, JOSEPHUS, AGRIPPA II AND HIS COINS The Literary Eras of Agrippa II A careful study of Josephus (with other relevant Greek and Latin sources) concerning the gradual accumulation of territories by Agrippa II reveals five potential ‘eras’ — de­ scribed here as ‘literary’, and named as those o f Chalcis, Panias (et alib i), Tiberias, Neronias and Area.35 When Herod of Chalcis died, since his son (Aristobulus III) was too young, Claudius temporarily assigned his kingdom to his nephew Agrippa II who lived in Rome. Herod’s death is clearly dated to Claudius’ eighth year (A nt. 20.104; cf. W ar 2. 223), that is between 25 January 48 and 24 January 49, and so probably he died sometime in 48. Tacitus (Ann. 12.23) would have helped to determine the period more precisely, if only he was not to confuse the identity o f Herod with that of his brother Agrippa I. But it may be indicative that he wrongly placed Agrippa’s death (44) early in the consulship of Q. Veranius and C. Pompeius (49), in which case the very end of Claudius’ eighth year is signaled for Herod’s death — that is between 1 and 24 January 49. Whether Herod died some time in 48 or at the beginning o f 49, there is no reason to assume any delay in the appointment of Agrippa II, which must have taken place by the early part of 49. In any case, given the Hellenistic, autumnal (‘Tishri’) reckoning of reg­ nal years followed by the Herodian dynasty,36 Agrippa’s first year will have been back­ dated, thus 48/9 (his second year will have begun in the autumn of 49). This is confirmed by A n t. 20.138 which specifies that this era lasted ‘four years’ (ἔ τη τέ σσαρα): that is 48/9, 49/50, 50/1, and 51/2. What is important to understand about this minor ‘era’, is that Agrippa’s reign over tiny Chalcis was basically in a bsen tia . He was still in Rome in c. 51/2, and apparently until the beginning o f 53 when Claudius switched his kingdom.37 ‘After the completion of his twelfth year’ (Ant. 20.138; cf. W ar 2.421), Claudius de­ prived Agrippa II of Chalcis and granted him instead the larger domain of Panias (with Abilene, Batanaea, Trachonitis and Auranitis attached). Since Claudius’ twelfth year was completed on 24 January 53, the appointment must have taken place after the end of January. Agrippa would now have moved from Rome into his new and considerable kingdom, where indeed we find him at the end o f 54 under Nero (Tac., A n n . 13.7). His ‘era’ of Panias (et a lib i ) will have been backdated — and so its first year will have been reckoned from 52/3 (matching the end o f the ‘era’ o f Chalcis). Now, Josephus (Ant. 20.159; cf. W ar 2.252) asserts that ‘in the first year of Nero’, Agrippa II received in addition portions of Galilee (principally Tiberias) and o f Peraea. Nero’s first year ran from 13 October 54 to 12 October 55. The event is grouped by Jo­ sephus with that of the appointment o f Aristobulus III, but mentioned subsequent in or­ der. We know from Tacitus that at the very end of 54 if not slightly later (A nn. 13.6 — f i n e a n n i), Aristobulus was assigned Armenia Minor (A nn. 13.7). Agrippa II’s grant would then fall within the year 55 and evidently many months on, as the grant must have awaited the result of Agrippa’s participation in the Parthian affair, which continued 35 36 For Agrippa II and his kingdom see, K H D 317-41. K H D 368-9 (Herod the Great), 234-5 (Antipas), 237 (Philip) 285 (Agrippa I — where refer­ 37 ences are given to a new consensus here; Stein, Meshorer and Burnett), 307 (Herod o f Chal­ cis), 398 (Agrippa II — now accepted by Kushnir-Stein 124). K H D 319-20 and notes 185-9. NIKOS KOKKINOS 173 for some time into 55 (A nn. 13.9 — q u a e in a lio s co n su les e g ressa co niu n xi). So although the first year o f Agrippa’s ‘era’ o f Tiberias may be thought to have been backdated to 54 (thus 54/5), the following year 55/6 cannot be excluded if the grant was received say as late as the end o f September 55 (and still within the first year of Nero which ended on 12 October). The period 59-60 saw the introduction by Nero in Rome o f many new entertainments, which began with the games called Iu v e n a lia (celebrating the first shaving of his beard) and ended with the N ero n ia (celebrating his ‘salvation’ and continuance of power). Capping the occasion, the N e ro n ia were instituted early in 60 as a quinquennial contest (Tac. A n n. 14.20; cf. Dio 61.19-21; Suet. N ero 11-2). Every prominent person was bound to contribute and the entrance of King Tiridates was included among the shows. Appropriately Agrippa II, who later in the year completed the enlargement of his capital Panias/Caesarea Philippi, renamed it ‘Neronias’ (Jos. A nt. 20.211). Thus the first year of the era of Neronias has to be 60/1, a fact independently supported by indisputable coin evidence (see below).38 The last addition to Agrippa II’s kingdom came some time after the final victory against the Jews in 73 (as reflected on his victory coins o f 73/4), and the occasion seems to have been the official celebration at the completion of the Templum Pacis in the be­ ginning of 75. Agrippa participated with his sister Berenice II, and he was also awarded the rank of p r a e to r (Dio 66.15). The additional territory concerned was almost certainly that of Area (Arcea), previously the tetrarchy of Varus who had been working for Ag­ rippa II but had been deprived of his position during the war (L ife 61, 180). In the usual conflating manner o f Josephus in the W ar (2.247), this award is placed together with that o f 53 — but this is obviously not accurate. In describing Agrippa’s kingdom at the beginning of the revolt (in the mid-60s), Josephus clearly omits Area (W a r 3.57), which he mentions retrospectively only after the end of the war (W a r 7.97). Decisively Justus, as we saw in B ib lio th e c a 's codex 33, testified that the last addition to Agrippa II’s kingdom was made by Vespasian (καἱ ἔ τι μ ἀ λλον ὑ πὸ Οΰεσπασιανοῦ ). Therefore the last ‘era’ o f Agrippa will have been backdated to 74, making its first year 74/5. It may be significant that in Eusebius’ C h ron ico n (ed. Helm, 179) the o n ly era of Agrippa II mentioned has ‘26’ years. This is arguably the la st era, since all other eras exceed this number, and they cannot serve the context. We know that Eusebius partly followed the Justus-Julius Africanus tradition. As it happens, Agrippa’s ‘twenty-sixth’ year according to his ‘era’ of Area is 99/100, the year in which his death was registered by Justus. Without numismatic and epigraphical input, the literary evidence shows that there were five potential ‘eras’ for Agrippa II: (1) Chalcis beginning in 48/9; (2) Panias (et a lib i) in 52/3; (3) Tiberias in 54/5 or 55/6; (4) Neronias in 60/1; (5) Area in 74/5 (see Table for the runs of these eras). My dating of this era beginning in 60 (previously thought to be 61 on coin evidence) has been accepted by Kushnir-Stein (123, n. 2). What has not yet been realised is that since the re-foundation o f Panias was carried out under Albinus (Ant. 20.197 & 211), the procurator must have arrived in Judaea in c. 59/60 — not in 62 (S V M 1: 468, n. 50); for the re-dating consequences including the execution of James the brother of Jesus, see K H D 385-6. Kushnir-Stein will not be able to dispute this case, for in her scheme the coin mentioning ‘Neronias’ would date to 59/60 (see below). 174 JUSTUS, JOSEPHUS, AGRIPPA II AND HIS COINS The Coin Eras of Agrippa II The bilingual (Latin-Greek) coins of Agrippa II, bearing his regnal years 25 and 26 to­ gether with the years of Domitian’s consulships Χ and XII, have finally established a major era beginning precisely in the autum n'of CE 60 (thus Year 1 = 60/1). My refine­ ment o f this era has now been accepted by Kushnir-Stein, and via her by Burnett et al. This era matches to perfection the era of Neronias (above no. 4). The refoundation of Agrippa II’s capital Panias/Caesarea Philippi as Neronias in 60 almost certainly was commemorated with a coin (undated) bearing the bust o f Nero and the reverse inscription ΕΠΙ BACIAEfOC] ΑΓΡΙΠΠ[Α] ΝΕΡΩ ΝΙΕ[Ω Ν].39 It was here that Agrippa retired after his expulsion from Jerusalem, and from here that he led the various opera­ tions during the war. It was also here that, in the presence o f Titus, he celebrated the victory of 70 with extravagant spectacles.40 Another coin of Panias (this one dated) of­ fers the opportunity to understand a second significant era o f Agrippa, and to deduce that the coin itself was minted in 65/6 upon his retirement to the city.41 The reading on the observe KAICAPIA TH ΚΑΙ ΝΕΡΩ ΝΙΑΔ Ι leaves no doubt about the place of origin. The date on the reverse explains how two (evidently major) eras o f Agrippa should be equated: ETOYC AI TOY ΚΑΙ f (‘of Year 11 which is also [Year] 6’). It would be absolutely incredible if one o f those two eras was not the era o f the coin’s own context, which era also happens to be the only one independently established by other coins — that o f Neronias. If we take Year 6 to represent the era of Neronias (60/1), Year 11 demands an era beginning in 55/6, and the one of Tiberias (above no. 3) provides a very reasonable match.42 That among the five literary eras, these two (Neronias and Tiberias) must have been the special ones is now clear. Agrippa II did not have a continuous kingship in one terri­ tory from beginning to end of his career. The Chalcis ‘era’ (or quasi-era in a b sen tia) was the most insignificant and was quickly put to one side. The Panias {et a lib i) ‘era’ was superseded by the refoundation o f Neronias, where Agrippan chronology became required as coins were issued there for the first time. The Tiberias ‘era’ was important to remember for it saw a significant move into the Galilee and Peraea, which brought the king closer to his ancestral lands. The ‘era’ of Area began too late to need to be counted universally. Although we have no other example of double dating on coins, we know that this system continued indefinitely, as it appears much later on two stone inscriptions: Year 21 = 16 of CE 75/6 {SE G 7.970), and Year 37 = 32 of CE 91/2 ( O G IS 426). This tells us that indeed only tw o major eras had been active during the reign of Agrippa II, and that the king wished to make this official. It also tells us that both of 39 40 41 42 T JC nos. 129-131 b; this coin was later countermarked with the initials o f the Tenth Roman Legion Fretensis — an example was found in Masada (no. 129a). K H D 326-8. T JC no. 132; cf. 133-133b for yet another type with the same date. The opposite, taking Year 11 as representing the era of Neronias, cannot work for it pro­ duces an era beginning in 65/6 which is not only unattested, but also conflicts with other evidence — e.g. the highest double date available in O G IS 426 would read 96/7, which is incompatible with the same date in IG R R 3.H 76 (see table in K H D 400). NIKOS KOKKINOS 175 these eras would have been used separately in one way or another, and that occasionally they had to be presented together to avoid confusion. Only in two cases can we be confident that the era o f Tiberias was used. Ἀ coin of Tiberias (TIBEPIAC) advertised the victory in Jerusalem (NIK[HC] CEB[ACTOY]) in Year 15 of Agrippa, that is precisely 69/70.43 There is no reason to imagine that Agrippa’s era of Neronias could or should have been imposed on the city o f Tiberias. At best, we might have hoped that the minting authorities at Tiberias would add ‘also Year 10’, returning some favour to the 65/6 coin o f Panias, but this was not to be, probably conveniently. The Tiberians were known for their superiority complex over other cities from the time of Antipas to Agrippa II (L ife 37-9). The second case is a lead weight from Tiberias (S E G 38, 1647), read by Qedar as of ‘Year 43’ and assigned by me to Agrippa II. In his era o f Tiberias, this translates into 97/8, which is an interestingly late date. Alas, even if irrelevant to the present point, Kushnir-Stein objected to Qedar’s reading and reportedly suggested ‘Year 33’ instead.44 She later seems to have changed her mind to ‘23’ (131, n. 32). Whatever the correct number, and we shall only know when Kushnir-Stein publishes a clear photograph, the weight belongs to Tiberias and thus Agrippa’s era of Tiberias must appropriately have been used. It is natural then to accept that the bulk o f Agrippa’s coinage, the Flavian coins, were minted at his capital Panias using his major local era o f 60/1. Indeed, following the es­ tablished chronology of the Latin-Greek coins mentioned, all available dates (14-15, 18­ 19, 24-27, 29-30, 34-35) can be accommodated within this single era. The result is that his latest known issue (35) circulated in 94/5, thus proving that the king was still alive beyond the assumed date for his death pre-93, and this seems to agree with Justus’ date of 100. The inscription of Archieus (A E 1966, 493) comes in conveniently here to suggest that Agrippa’s reign ran into that o f Trajan beginning in 98. Jones (116) states that ‘nothing indicates that the man [Archieus] went immediately from his [Agrippa’s] service to the emperor’s [Trajan]’, and he invites comparison with the revised Schürer (n. 14). This is interesting. Schürer clearly admitted that were it not for other reasons this inscription ‘... would undoubtedly suggest that the man’s [Archieus’] service under Trajan had succeeded that under Agrippa directly’.45 But Jones ignores the fact that even closer agreement exists. Tiberias actually declared the change of power at the death of its king by immediately issuing provincial coins (for the first time for half a century) — acknowledging Trajan in Year 81 dated from the city’s foundation in 19/20 under Antipas.46 This year is 99/100, and since Agrippa according to Justus ‘died in the third year o f Trajan’, running from 28 January 100 to 27 January 101, the king must have died between February and mid-September (when Year 82 o f Tiberias began). He 43 44 45 46 T JC no. 134. The legend follows the V ictoria A u g u sti already appearing on the coins of Vitellius in 69 and reflecting the first Roman victories in Judaea (B M C R E 1: 379, no. 61). See Mason (n. 8) xviii. S V M 1: 483, n. 47 (8). B M C Palestine, xiv; Y. Meshorer, C ity-C o in s o f E retz-Isra el a n d th e D e c a p o lis in the R o ­ m a n P e rio d (Jerusalem, 1985) 34, no. 77. The previous provincial coins o f Tiberias {TJC nos. 347-9a) were issued under Claudius in 53/4 (before Agrippa II) during the procuratorship o f Felix. Meshorer continues to misunderstand the status o f Galilee (Tiberias 44 to 55, and Sepphoris 44 to 67) — see K H D 234, n. 103; 320, n. 188; 398; cf. 289, n. 88. 176 JUSTUS, JOSEPHUS, AGRIPPA II AND HIS COINS was 72 years old.47 At this time Justus’ H isto ry o f the Jew ish W ar would have been published, instigating the revision of Josephus’ L ife. Josephus was 62 years old and evidently still strong and argumentative.'*8 Kushnir-Stein (124) has no choice but to agree with the era o f 60/1 at least for the years up to and including 25. She questions 26, 27, 29 and 30 — which she calls ‘anach­ ronistic’ — and she remains undecided about 34 and 35. As only the latter are relevant to the date of Agrippa’s death, Kushnir-Stein makes no contribution to the specific debate and thus these years can be left out here. Her reasons for questioning the middle years are not new. Certain anomalies in the iconography and inscriptions have been known for centuries. But once isolated they cannot outweigh the overall structure. Even in the period which Kushnir-Stein accepts, for example Year 14 = 73/4 and Year 18 = 77/8 of the Greek series, Titus is mentioned as ‘Augustus’, a title which he did not officially acquire before he became Emperor (79).49 An explanation for this anomaly would be that the title was not applied in the East as strictly as in the West, as I have proposed in a different context.50 Another example o f the Greek series is Year 24 = 83/4 and Year 25 = 84/5, where Domitian (Emperor in 81), appears without his full imperial titles, and not yet upgraded to the Tyche type but still using both his old Nike-shield and his brother’s old Nike-wreath.51 If it were not for the title ‘Germanicus’, these coins too could have been condemned as anachronistic by Kushnir-Stein. We should further note that Domitian appears without titles (only as ‘Caesar’) on the coins o f Canatha as late as 94/5 and 95/6.52 Although this is not to deny any anomalies in the middle years of Agrippa’s coins, some issues may well be exempt from anomalous status, urging caution in respect of Kushnir-Stein’s analysis. I pick three types. In Year 26 = 85/6 a specimen of Domitian with his old Nike-shield type includes his title ‘Germanicus’ which he gained in 83.53 This coin cannot be thrust back some twelve years to where Kushnir-Stein wants it (see below). In Year 27 = 86/7, Domitian is shown with his old Nike-shield type on the obverse, and crossed cornucopiae on the reverse.54 The motif o f a ‘double cornucopia’ is known to be associated with Domitian only on a few issues in the area, clustered in the period 82/3 to 86/7. This last is an example from Neapolis.55 It would seem out of place some twelve years earlier. In Year 29 = 88/9 a coin presents Domitian with full titles (including ‘Germanicus’) and for the 47 48 49 50 51 52 53 54 55 He was bom in CE 27/8 (K H D 276). Herod the Great died ‘nearly 70’ {K H D 156, n. 2). In Jones’ assumption for the death of Agrippa II in CE 88/89?, the king would have been 61. K H D 392 (e). In Jones’ assumption (121) for the death o f Josephus ‘probably not later than [CE] 96’, he would have been 58 or less. R P C 2: nos. 2246, 2255 = T JC nos. 137, 143; see T."V. Buttrey, D o c u m en ta ry E vid en ce f o r th e C h ro no log y o f th e F la v ia n T itulature (Meisenheim am Gian, 1980) 14. A n to n ia A u g u sta (London/New York, 1992) 103; cf. now 265-7 in the review chapter o f the paperback edition (London, 2002). I am grateful to Theodore Buttrey for discussing whether the usage in the provinces was sometimes non-conforming, either due to ignorance or be­ cause o f boot-licking. R P C 2: nos. 2262, 2264; TJC no. 155. C D PA 92-3, nos. 4-5. T JC no. 165c = R P C 2: no. 2281; see Buttrey (n. 49) 53-6. T JC no. 171 = R P C 2: no. 2287. R P C 2: nos. 2219 (82/3), 2224 (86/7), 2268 (84/5), 2271 (85/6). NIKOS KOKKINOS 177 first time in the Greek series upgraded to the Tyche type.56 It is absolutely impossible to move this issue back in time before 83. O f course, Kushnir-Stein attempts to diminish the value of such specimens (e.g. 125, n. 8), but she cannot argue against their existence. Most of the anomalous coins, nevertheless, are anomalous. Yet it must be made clear that there is an adequate explanation for the phenomenon. The Flavian period in Rome saw many ‘restorations’ o f older issues, and the trend or economic necessity would have affected the only important client kingdom remaining in the eastern part of the Empire, that o f Agrippa II. Mattingly described it best: ‘The restorations o f types o f a es by Titus, Domitian and Nerva, certainly suggest that the types restored were already going out of circulation. The reduction of the weight of aureus and denarius by Nero sentenced the bulk of the earlier gold and silver to the melting-pot; that is, perhaps, in fact, the reason why Vespasian and Titus more or less closely copy a number o f the earlier types’.57 Is it therefore not preferable to live with this explanation rather than to reshuffle the entire numismatic evidence and play table-games with it? To understand the difficulties presented by Kushnir-Stein’s hypothesis, the concept o f her scheme must be outlined. She attempted (127, Table 3) to explain the anomalies in years 26, 27, 29 and 30, by applying a different era to these coins — that is an earlier one which would make them more or less contemporary with years 14 to 19 o f the established era of 60/1. The only era which might have done the job is the era o f Chalcis 48/9. But Kushnir-Stein would have realized that making a coin o f Year 30 equivalent to 77/8, which upgrades Titus to the Tyche type (reserved thus far for the Emperor),58 would have heralded his accession (79) one year too early. So she moved this era a year later to 49/50. No evidence was presented beyond the statement (127) that ‘Seyrig has shown that the starting point... is the autumn of 49’. However Seyrig’s argument was merely circular.59 Josephus (W a r 2.284) in a literary calculation (with no consideration o f coin eras and without ever using it again throughout his writings) gives the ‘seventeenth’ year from the beginning of Agrippa’s kingship as the time o f the beginning of the Jewish Revolt. Since the latter is assumed to have started in 66 — based on unproven emendations — Agrippa’s reign over Chalcis would have begun in 4 9 /5 0 .1 have already reversed this argument (ridding it o f circularity) to show that since the era of Chalcis started in 48/9 (with no emendation), the Revolt would have begun in 65.60 In fact, Seyrig’s article should not have been invoked at all, for it failed on cardinal points including missing the basic date o f Domitian’s becoming ‘Germanicus’.61 56 57 58 59 60 61 T JC no. 174 = R P C 2: no. 2289. Tyche o f course had appeared with Domitian outside Agrippa’s kingdom already in 82/3 at Pella {C D PA 212-3, no. 2). B M C R E lixxii, n. 1. T JC no. 176 = R P C 2: no. 2294. Η. Seyrig, ‘Les ères d’Agrippa’, R N 5/6 (1963/4) 56. K H D 391 (a). See discussion in Schwartz (n. 5) 259-61. Even if we were to grant Kushnir-Stein an era of 49/50, her scheme is tight for Agrippa’s Year 30 coin with Titus and Tyche (her 78/9). Titus became Emperor on 24 June. In mid-September Agrippa’s reign would enter Year 31. This leaves some 80 days for the news to reach Syria, and for Agrippa to order new coins before his new year was celebrated. Among irregularities would be the amount o f coins minted for Year 27 (her 75/6) which would parallel ‘Year 16’ of the original series o f 60/1 — an 178 JUSTUS, JOSEPHUS, AGRIPPA II AND HIS COINS Satisfied with her new era, and since Agrippa’s brief affair with Chalcis had long been forgotten (the kingdom was currently in Aristobulus Ill’s hands), Kushnir-Stein now found herself obliged to discover a new home for it. As the anomalous coins were grouped separately, one thinks that a minting place would have been sought away from Panias (where the normal coins will have been minted). But against expectations, another surprising suggestion was made by Kushnir-Stein. In Year 27 = 86/7 (but 75/6 in her scheme) a coin depicts, anomalously, Titus and Domitian face-to-face, with its reverse illustrating a statue o f Pan.62 There can be no doubt that this coin (praising the local deity) was minted at Panias. If the explanation for the anomaly is that it is a ‘restored’ coin (even if an ‘original’ example has yet to be unearthed), then more evidence is provided for the era concerned (being linked to Panias as it is) actually being the established one of ‘Neronias’ followed by all other coins. But for Kushnir-Stein, on the contrary, this had to mean that the anomalous coins themselves were minted at Panias, and thus all o f the normal coins following the established 60/1 era needed to find a new home! Here the scheme begins to run off the rails (128-9) as follows. In 60/1 Agrippa II m a y have acquired a new territory, conceivably o f g r e a t importance yet previously u n reco rded , where all of his undisputed Flavian coins would have been minted. No evidence is presented as to where this might be, and no provenanced coins are produced in support of the claim. To Panias, even though it was Agrippa’s capital city, are to be attributed only the disputed ones, which began to be struck only in 74/5, and only until 78/9 (or 83/4?). We are thus asked to believe in a sudden introduction, and an equally sudden disappearance, o f a minor and ‘antiquated’ era from Chalcis in preference to that o f Panias in Panias — without a clue as to what Panias itself was minting before or after, and while the major ‘Neronias’ era o f 60/1 continued to be active away from the capital and in an unknown location. Any attempt to rationalize such a scenario should ask how anybody was expected to guess which era is meant on the coins attributed by Kushnir-Stein to the era of 49 — paralleled co n tem p o ra n eo u sly by issues dated by the era o f 60? There is no sign on the coins that a different era is meant or that the issue comes from a different mint. Years ago, Smallwood had already felt very uncomfortable with the application o f the Chalcis era on the coins of Agrippa II: ‘... the use of that out-of-date era of Agrippa is unexam­ pled and seems highly improbable’.63 But the scheme’s derailment continues much further. What about the known double dating of Agrippa, separating two eras by six years (inclusively)? The era of Chalcis is unfortunately 12 years (in Kushnir-Stein’s counting from 49/50, but in fact 13 counting from 48/9) away from that of 60/1. We are told that this should not be a problem. The double dating simply does not involve the main, established era of 60/1 in whichever territory it is being isolated. Another era must have existed six years later than Chalcis. Such an era would be starting in 54/5 (in fact 53/4), and although the era o f Tiberias might just have qualified, its accommodation should not even be attempted. Why? It 62 63 unattested year. One of these coins is that of Domitian with the double cornucopia (mentioned above as known only in the mid-80s). TJC no. 168 = R P C 2: no. 2284. Μ. Smallwood, The J e w s U nder R o m a n R u le (Leiden, 1981) 574. I am grateful to Dan Barag for discussion on this point. NIKOS KOKKINOS 179 suits the scheme better boldly to move the era o f Panias {et a lib i) from 53/4 (in fact 52/3) down to 54/5 and there it is. Without any mention o f Tiberias, it can thus be explained how the coin bearing the double date ‘Year 11 which is also [Year] 6’ (= 65/6 mentioned above) was minted at Panias in good time, 59/60, for the city to have been called ‘Neronias’ (cf. above note 38). And what about the coin from Tiberias advertising the victory in Jerusalem in Agrippa’s ‘Year 15’ (= 69/70 mentioned above)? Unfortunately in this case the new double era is too early for it. But all is not lost. The ‘expelled’ era o f 60/1 can still be called upon from its unknown territory and be imposed on Tiberias. This will at least give a delayed commemoration in 74/5, which is not too late considering that ‘victory’ coins began to be issued by Agrippa only in 73/4 in agreement with Masada’s fall in 73.64 We could certainly do without a commemoration by Agrippa precisely in 69/70 (even if he participated in the war and even if Victoria A u g u sti coins appeared in Rome already in 69), because this will only introduce yet another era — that o f Tiberias (where the coin was actually struck). The scheme simply cannot afford more eras; from the previous one to two eras, we have now gone up to three, and a fourth one will surely stretch our imagination (given that Agrippa’s coins and inscriptions inform us only of a double era). Therefore three eras are being proposed by Kushnir-Stein: a distorted era o f Chalcis (49/50 instead o f 48/9) and a distorted era of Panias et a lib i (54/5 instead o f 52/3), which together form a double era imposed on all territories, but not on an unrecorded territory where the established era of 60/1 (thought to have risen out o f the re­ foundation o f Panias as ‘Neronias’ exactly in 60) was active; this era, nevertheless, can be imposed on Tiberias whenever need be (as in the case o f the victory coin o f Year 15). But I must rest my case here. Kushnir-Stein’s scheme is speculative and lacks positive evidence. It appears to solve one problem but introduces new problems which render her suggestion unconvincing. To sum up, we can now see that there is no conflict between Justus and Josephus on the date o f Agrippa II’s death — whatever the consequences are for the relationship, date o f publication, and re-edition o f the A n tiq u ities and L ife. The latest coin o f the king dates to 94/5, beyond the year Josephus specifies for the completion (not the publication) o f the A n tiq u itie s (not the L ife). The inscription of Archieus shows that it is reasonable to believe that the king’s reign ran into that of Trajan beginning in 98. The city of Tiberias duly acknowledged the fact of its submission to Trajan, by issuing new coins for the first time in fifty years precisely in 100. Justus was right all along: Agrippa II ‘died in the third year o f Trajan’. London University 64 For the date of the fall o f Masada see K H D 199, n. 98 (in agreement with Bowersock); the Year 14 coins of Agrippa correctly dated to 73/4 (not 74/5) answer the interesting questions raised by D. Barag in M a sa d a I V (Jerusalem, 1994) 3, n. Ι (Α). 180 JUSTUS, JOSEPHUS, AGRIPPA II AND HIS COINS Table: The Eras of Agrippa II 48/49 = 01 — Chalcis 49/50 = 02 50/51 = 03 51/52 = 04 52/53 = 05 = 01 — Panias {et alib i) 53/54 = 06 = 02 54/55 = 07 = 03 55/56 = 08 = 04 = 01 — Tiberias 56/57 = 09 = 05 = 02 57/58= 10 = 06 = 03 58/59 = 11 = 07 = 04 59/60= 12 = 08 = 05 60/61 = 13 = 09 = 06 = 01 — Neronias 61/62= 14= 10 = 07 = 02 62/63 = 15 = 11 = 0 8 = 03 63/64= 16= 12 = 09 = 04 64/65 = 17= 13= 10 = 05 65/66= 18 = 14 = 11 = 06 66/67= 19= 15 = 12 = 07 67/68 = 2 0 = 16= 13 = 08 68/69 = 21 = 17== 14 = 09 69/70 = 2 2 = 18= 15 = 10 70/71 = 23 = 19= 16 = 11 71/72 = 24 = 2 0 = 17= 12 72/73 = 25 = 21 = 18 = 13 73/74 = 26 = 2 2 = 19= 14 74/75 = 27 = 23 = 20 = 15 = 01 — Arca 75/76 = 28 = 24 = 21 = 16 = 02 76/77 = 29 = 25 = 2 2 = 17 = 03 77/78 = 30 = 26 = 23 = 18 = 04 78/79 = 31 = 2 7 = 2 4 = 19 = 05 79/80 = 32 = 28 = 25 = 20 = 06 80/81 = 33 = 29 = 26 = 21 = 07 81/82 = 34 = 30 = 27 = 22 = 08 82/83 = 35 = 31 = 2 8 = 23 = 09 83/84 = 36 = 32 = 29 = 2 4 = 10 84/85 = 37 = 33 = 30 = 25 = 11 85/86 = 38 = 34 = 31 = 2 6 = 12 86/87 = 39 = 35 = 32 = 2 7 = 13 87/88 = 40 = 36 = 33 = 28 = 14 88/89 = 41 = 3 7 = 34 = 2 9 = 15 89/90 = 42 = 38 = 35 = 3 0 = 16 90/91 = 43 = 39 = 36 = 31 = 17 91/92 = 44 = 40 = 37 = 3 2 = 18 92/93 = 45 = 41 = 3 8 = 3 3 = 19 93/94 = 46 = 42 = 39 = 34 = 20 94/95 = 47 = 43 = 40 = 35 = 21 95/96 = 48 = 44 = 41 = 3 6 = 22 96/97 = 49 = 45 = 42 = 37 = 23 97/98 = 50 = 46 = 43 = 38 = 24 98/99 = 51 = 4 7 = 44 = 39 = 25 99/100 = 52 = 48 = 45 = 4 0 = 26